LUFS, dBFS and True Peak: What Each Meter Tells You

Production guides · Meme Plugins · 26 October 2026

LUFS describes measured programme loudness, a sample-peak reading in dBFS shows the highest digital sample level, and true peak estimates the peak of the reconstructed waveform in dBTP. Read them together: one number cannot tell you both how loud a mix feels over time and whether its output has enough peak margin.

This guide explains what each reading can answer. For a practical listening comparison, use the separate reference-track workflow.

dBFS: how high are the samples?

dBFS means decibels relative to digital full scale. On a sample-peak meter, a reading of −3 dBFS means the largest measured sample is three decibels below full scale. In an integer PCM output file, 0 dBFS is the full-scale limit. This describes the samples, not the average loudness or the sound pressure reaching your ears.

A single drum hit can produce a high sample peak while most of the track remains quiet. The peak alone does not explain the density of the chorus. FabFilter’s metering documentation distinguishes peak readings and metering scales; check which mode your own meter is displaying.

dBTP: what can happen between samples?

The reconstructed waveform can peak above the individual sample values. A true-peak meter estimates that behavior and reports dBTP. A file can therefore have sample peaks below 0 dBFS while its true-peak estimate is higher.

FabFilter explains inter-sample peaks, including peaks that can appear during lossy conversion. True-peak measurement is useful when checking output margin, but one reading does not predict every encoder, playback device or listening condition. Follow the actual delivery specification and inspect the encoded file when it matters.

LUFS: how loud is the programme over time?

LUFS means loudness units relative to full scale. Integrated programme loudness uses frequency weighting and gating rather than simply selecting the tallest sample. The Audio Engineering Society’s loudness introduction explains how the method relates to perceived loudness. It remains a signal measurement, not a guarantee that every listener hears two sounds identically.

  • Momentary: follows a short window and responds quickly.
  • Short-term: shows a longer moving window, useful for observing a passage.
  • Integrated: summarizes the material measured since the measurement began or was reset.

FabFilter’s loudness-meter guide describes these time scales. Measure the complete intended programme for a whole-track integrated result; do not compare one file’s chorus with another file’s full-song reading.

Example: the same peak, different loudness

Imagine two versions of a song that both reach −1 dBFS sample peak. Version A has isolated drum transients and open space. Version B has a dense, sustained arrangement and more limiting. Hypothetical integrated readings of −17 LUFS and −11 LUFS would not contradict their identical sample peaks. The second version has the higher measured programme loudness.

Those figures are an illustration, not measurements of supplied audio or mastering targets. Peak-normalizing both versions to the same value would not by itself make them equally loud.

A three-question meter check

  1. Am I reading the intended thing? Note the units and whether the meter is showing sample peak, true peak or a loudness time scale.
  2. Did I measure comparable material? Use the full export for an integrated check, then inspect the loudest sections separately.
  3. What decision follows? Investigate excessive peaks, compare loudness fairly, and listen for unwanted distortion. Do not change the arrangement simply to make every display look similar.

The Reckoning provides a browser-based starting point for inspecting a mix. Treat its results as context for listening. The underlying measurement subject is covered by ITU-R BS.1770; a meter’s presence alone does not establish certification or identical implementation.

Meter questions

Does a lower LUFS number mean lower quality?

No. A more negative reading means lower measured loudness. Musical dynamics, distortion, tonal balance and the intended delivery all need separate judgment.

Does turning down the speakers change LUFS?

Changing only your listening volume does not change the audio file’s measured loudness. Changing gain in the signal being measured does.

Is there one LUFS target for every song?

No. Delivery requirements and playback behavior vary. Use the requirements for the destination and make a musical choice; do not treat a platform playback reference as a compulsory creative target.

If the problem is unwanted overload, continue with how to fix clipping in an audio mix. Browse the production guides for more workflows.


Browse all production guides → · Free browser audio tools